Controlling the Biological Effects of Spermine Using aSynthetic Receptor

Laurent Vial,† R. Frederick Ludlow, Julien Leclaire,‡ Ruth Perez-Fernandez, andSijbren Otto*

Contribution from the Department of Chemistry, UniVersity of Cambridge, Lensfield Road,Cambridge CB2 1EW, United Kingdom

Received April 12, 2006; E-mail: so230@cam.ac.uk

Abstract: Polyamines play an important role in biology, yet their exact function in many processes is poorlyunderstood. Artificial host molecules capable of sequestering polyamines could be useful tools for studyingtheir cellular function. However, designing synthetic receptors with affinities sufficient to compete withbiological polyamine receptors remains a huge challenge. Binding affinities of synthetic hosts are typicallyseparated by a gap of several orders of magnitude from those of biomolecules. We now report that adynamic combinatorial selection approach can deliver a synthetic receptor that bridges this gap. The selectedreceptor binds spermine with a dissociation constant of 22 nM, sufficient to remove it from its natural hostDNA and reverse some of the biological effects of spermine on the nucleic acid. In low concentrations,spermine induces the formation of left-handed DNA, but upon addition of our receptor, the DNA revertsback to its right-handed form. NMR studies and computer simulations suggest that the spermine complexhas the form of a pseudo-rotaxane. The spermine receptor is a promising lead for the development oftherapeutics or molecular probes for elucidating spermine’s role in cell biology.

Introduction

Spermine (1, Figure 1) is a polycation that plays a key rolein many cellular processes1,2 forming strong interactions withDNA and RNA.1,3 Spermine has been shown to condenseDNA3,4 and chromatin,5,6 to facilitate the formation of nucleo-somes,7 to induce8-10 and to stabilize8,9 specific DNA confor-mations, to modulate DNA-protein interactions,11 and to protectDNA from external agents.12 These phenomena may underliespermine’s ability to regulate cell proliferation1 and influencetumor growth13,14 and apoptosis.15 The exact mechanism of

action of spermine in most of these processes is still underdebate. Thus, there is a considerable incentive to developsynthetic molecules that can sequester spermine under physio-logical conditions. While receptors have been reported,16,17

designing a water-soluble synthetic receptor with sufficientaffinity to interact with spermine in biological systems remainsa huge challenge. Binding affinities of synthetic hosts aretypically in the millimolar range, while biological systemsusually exhibit submicromolar affinities.18 We now report thata dynamic combinatorial selection approach19-21 can deliver asynthetic receptor that bridges this gap and binds spermine withnanomolar affinity, sufficient to remove it from its biologicalhost DNA in vitro, causing the nucleic acid to change itsconformation from a left-handed to a right-handed helix. NMRstudies and computer simulations suggest the spermine complexhas the form of a pseudo-rotaxane.

Results and Discussion

As spermine lacks a well-defined geometry in solution,designing a receptor for it is inherently difficult. We thereforetook a selection approach, generating mixtures of potentialreceptors by linking building blocks together using a reversiblereaction. The resulting dynamic combinatorial libraries (DCLs)

† Current address: LEDSS, UMR 5616 CNRS- Universite JosephFourier, 301 Rue de la Chimie, BP 53, 38041 Grenoble cedex 9, France.

‡ Current address: UMR 6180, Case A62, EGIM- Sud, Faculte´ deSaint-Je´rome, Av. Escadrille Normandie-Nie´men, 13397 Marseille cedex20, France.(1) Tabor, C. W.; Tabor, H.Annu. ReV. Biochem.1984, 53, 749-790.(2) Igarashi, K.; Kashiwagi, K.Biochem. Biophys. Res. Commun.2000, 271,

559-564.(3) Bloomfield, V. A. Biopolymers1997, 44, 269-282.(4) Flock, S.; Labarbe, R.; Houssier, C.Biophys. J.1996, 70, 1456-1465.(5) Hougaard, D. M.; Bolund, L.; Fujiwara, K.; Larsson, L. I.Eur. J. Cell

Biol. 1987, 44, 151-155.(6) Smirnov, I. V.; Dimitrov, S. I.; Makarov, V. L.J. Biomol. Struct. Dyn.

1988, 5, 1149-1161.(7) Basu, H. S.; Smirnov, I. V.; Peng, H. F.; Tiffany, K.; Jackson, V.Eur. J.

Biochem.1997, 243, 247-258.(8) Basu, H. S.; Schwietert, H. C. A.; Feuerstein, B. G.; Marton, L. J.Biochem.

J. 1990, 269, 329-334.(9) Garriga, P.; Garcia-Quintana, D.; Sagi, J.; Manyosa, J.Biochemistry1993,

32, 1067-1071.(10) Thomas, T. J.; Messner, R. P.Nucleic Acids Res.1986, 14, 6721-6733.(11) Oller, A. R.; Vandenbroek, W.; Conrad, M.; Topal, M. D.Biochemistry

1991, 30, 2543-2549.(12) Ha, H. C.; Yager, J. D.; Woster, P. A.; Casero, R. A.Biochem. Biophys.

Res. Commun.1998, 244, 298-303.(13) Thomas, T.; Thomas, T. J.J. Cell. Mol. Med.2003, 7, 113-126.(14) Khuhawar, M. Y.; Qureshi, G. A.J. Chromatogr., B2001, 764, 385-407.(15) Seiler, N.; Raul, F.J. Cell. Mol. Med.2005, 9, 623-642.

(16) Isobe, H.; Tomita, N.; Lee, J. W.; Kim, H. J.; Kim, K.; Nakamura, E.Angew.Chem., Int. Ed.2000, 39, 4257-4260.

(17) Jeon, Y. M.; Whang, D.; Kim, J.; Kim, K.Chem. Lett.1996, 503-504.(18) Houk, K. N.; Leach, A. G.; Kim, S. P.; Zhang, X. Y.Angew. Chem., Int.

Ed. 2003, 42, 4872-4897.(19) Ramstro¨m, O.; Lehn, J.-M.Nat. ReV. Drug DiscoVery 2002, 1, 26.(20) Otto, S.Curr. Opin. Drug DiscoVery DeV. 2003, 6, 509-520.(21) Rowan, S. J.; Cantrill, S. J.; Cousins, G. R. L.; Sanders, J. K. M.; Stoddart,

J. F.Angew. Chem., Int. Ed.2002, 41, 899-952.

Published on Web 07/13/2006

10.1021/ja062536b CCC: $33.50 © 2006 American Chemical Society J. AM. CHEM. SOC. 2006 , 128, 10253-10257 9 10253

are equilibrium mixtures that are responsive to influences thatresult in the selective stabilization of specific library members.Thus, introducing a guest into a DCL of potential hosts willshift the equilibrium, amplifying strong binders at the expenseof their inferior counterparts.20,21 Amplification allows newsynthetic receptors to be identified and, in small DCLs, evenprovides a method for their preparation.22-24

We have set up a DCL from building blocks225 and 3equipped with carboxylate groups (Figure 1) tailored to recog-nize the protonated amine groups of spermine through hydrogen-bonding and cation-anion interactions. We deliberately intro-duced a large number of carboxylate groups as each amine groupof spermine is able to interact with several of these units. Anexcess of carboxylate groups also ensures the solubility of thelibrary members in water by preventing neutralization of allnegative charges by the cationic spermine. The thiol functional-ity allows the building blocks to be oxidatively linked togetherto form disulfide bonds, which can subsequently undergodisulfide exchange to generate the DCL.26

As it was not clear whether linear or cyclic architectureswould be more suitable for spermine recognition, we designedour libraries to contain both architectures (Figure 1). Thus, whilethe oxidation of dithiol2 will give macrocyclic species ofvarious ring sizes, the presence of monothiol3 will enable theadditional formation of linear species.

Equimolar amounts of the building blocks2 and 3 wereallowed to oxidize and equilibrate in aqueous solution at pH8.25 in the presence of air following standard protocols.26 Thecomposition of the ensuing DCL was analyzed by HPLC andmass spectrometry (Figure 2a). Linear tetramer4 was the majorcompound in the absence of spermine. When spermine wasintroduced into the DCL, the cyclic tetramer6 was amplified,alongside linear dimer5 built up from two units of monothiolbuilding block 3. The strong affinity of spermine for6 (videinfra) drives the conversion of dithiol2 into this receptor andleaves monothiol3 with no alternative but to dimerize.

A biased DCL made from only building block2 and sperminegives macrocycle6 in more than 90% yield allowing straight-forward isolation by preparative HPLC (Figure 2c and d).

The structure of the complex between6 and spermine wasinvestigated using proton NMR spectroscopy and computermodeling.

Upon binding of spermine to6 a strong ring-current-inducedupfield shift was observed for the central protons d and e ofspermine, while protons a, b, and c near the termini of themolecule are hardly shifted (Figure 3). These observationssuggest that the central fragment of spermine is located withinthe cavity of the receptor and imply the formation of a pseudo-rotaxane in which spermine is threaded through the cyclicreceptor.16,17

The presence of multiple signals for the host-guest complexis most likely a result of stereoisomers resulting from differentconformations of the disulfide linkages. The preferred 90°torsional angle of the disulfide bond makes this linkageinherently chiral as it can exist in a P or Mconfiguration. Inmost disulfides, rapid rotation around the S-S bond leads tofast exchange between stereoisomers. However, when spermineis threaded through macrocycle6, this rotation is likely to behindered, thus freezing out up to four different diastereomersthat arise from the fact that the four disulfide bonds can havedifferent configurations. The NMR spectrum of the complex isdominated by a singlet for the host protons, indicating that the

(22) Otto, S.; Furlan, R. L. E.; Sanders, J. K. M.Science2002, 297, 590-593.(23) Lam, R. T. S.; Belenguer, A.; Roberts, S. L.; Naumann, C.; Jarrosson, T.;

Otto, S.; Sanders, J. K. M.Science2005, 308, 667-669.(24) Corbett, P. T.; Tong, L. H.; Sanders, J. K. M.; Otto, S.J. Am. Chem. Soc.

2005, 127, 8902-8903.(25) Field, L.; Engelhardt, P. R.J. Org. Chem.1970, 35, 3647-3655.(26) Otto, S.; Furlan, R. L. E.; Sanders, J. K. M.J. Am. Chem. Soc.2000, 122,

12063-12064.

Figure 1. Oxidation of thiol building blocks2 and3 produces a DCL of linear and macrocyclic disulfides that contains potential receptors for protonatedspermine (1).

Figure 2. HPLC analyses of the libraries made from thiols2 and3 (10mM each) in the absence of any template (a), and in the presence of spermine(2.5 mM) (b). HPLC analyses of the libraries made from thiol2 only (10mM) in the absence of any template (c),27 and in the presence of spermine(2.5 mM) (d).

A R T I C L E S Vial et al.

10254 J. AM. CHEM. SOC. 9 VOL. 128, NO. 31, 2006

major stereoisomer is highly symmetric with all four disulfidebonds having the same configuration.

Computer modeling studies on the free receptor6 and on itscomplex with spermine provide further support for theseconclusions. In the absence of spermine, the four diastereomersof the receptor were found to be similar in energy. Modelingof the four different spermine-receptor complexes showed thatthe diastereomer with the highest symmetry was significantlylower in energy than the three others. Furthermore, the energybarrier for the interconversion of the diastereomers was foundto be substantially higher in the spermine complex than in thefree receptor (see Supporting Information).

In the lowest-energy conformation of the major pseudoro-taxane complex, 8 of the 10 amine protons of the guest areinvolved in hydrogen bonds with the host (Figure 4). In solution,rapid exchange between different hydrogen-bonding arrange-ments must occur as any long-lived hydrogen bonds woulddesymmetrize the receptor and no such effect is observed inthe proton NMR analysis.

The dissociation constant for binding of spermine to receptor6 was found to be 22 ((1) nM, as determined by isothermaltitration calorimetry (ITC) (see Supporting Information). Thisaffinity is higher than that reported for the binding of spermineto DNA,8,28-30 suggesting that6 should be able to inhibit DNA-spermine interactions.

Binding experiments were performed in the presence of DNAto assess the ability of the receptor to (i) inhibit the formationof DNA-spermine complexes, (ii) unbind spermine from DNA,and (iii) control the conformation of DNA.

When an excess of spermine (2.2 equiv with respect to thenumber of DNA base pairs) was added to a solution of calfthymus DNA, essentially complete precipitation of the biopoly-mer was observed as was apparent from the disappearance ofthe UV absorbance of the DNA (Figure 5a, red and blue traces,respectively). However, in the presence of6 (1.0 equiv withrespect to spermine), the addition of spermine does not giveany precipitation of DNA and the UV absorbance is unchanged(green trace).

At lower concentrations, spermine binding can switch theconformation of DNA from a right-handed (B-form) to a left-handed (Z-form) helix, a change that can be monitored bycircular dichroism (CD).8,10 Thus, the addition of one-half anequivalent of spermine (with respect to the number of DNAbase pairs) to a buffered solution of poly(dA-dC)‚poly(dG-dT)induced the appearance of a negative CD-band at 294 nm anda shift in the positive band from 280 to 263 nm (Figure 5b).

(27) The complexity of the trace can be explained by the presence of peakscorresponding to linear species in which the terminal thiols have oxidizedto sulfinic or sulfonic acids.

(28) Morgan, J. E.; Blankenship, J. W.; Matthews, H. R.Arch. Biochem. Biophys.1986, 246, 225-232.

(29) Ouameur, A. A.; Tajmir-Riahi, H. A.J. Biol. Chem.2004, 279, 42041-42054.

(30) Patel, M. M.; Anchordoquy, T. J.Biophys. J.2005, 88, 2089-2103.

Figure 3. 1H NMR (500 MHz, cryoprobe) analyses of receptor6 (a),spermine (b), and receptor6 (1.3 mM) in the presence of 1 equiv of spermine(c). Spectra were recorded at 300 K in 89 mM phosphate buffer (pH 7.4,11% D2O) with water presaturation.

Figure 4. Side and top views of the energy minimized complex of receptor6 and spermine obtained by computer modeling.32 Some hydrogen atomshave been omitted for clarity.

Figure 5. (a) UV spectra of calf thymus DNA (33µM in 3.0 mM TRISbuffer; pH 7.4; 298 K) without additive (blueO); in the presence of spermine(2.2 equiv with respect to the number of base pairs) (red]); and in thepresence of spermine and6 (2.2 equiv each) (green4). (B) CD spectra ofpoly(dA-dC)‚poly(dG-dT) (97µM in 3.0 mM TRIS/EDTA buffer; pH 7.4;1.0 mM NaCl; 298 K) without additive (blueO); after addition of spermine(0.5 equiv with respect to the number of base pairs) (red]); and afteraddition of 6 (0.5 equiv) (green4). The UV and CD traces correspondexclusively to DNA; contributions from the buffer, spermine, and6 havebeen subtracted.

Controlling the Biological Effects of Spermine A R T I C L E S

J. AM. CHEM. SOC. 9 VOL. 128, NO. 31, 2006 10255

Subsequent addition of one-half an equivalent of6 to theresulting Z-DNA solution reversed the transition, regeneratingthe original B-DNA, indicating that the receptor is indeed ableto unbind spermine from DNA and thereby control the helicityof DNA.31

Conclusions

Our results demonstrate that, after the design of onlyrudimentary molecular recognition elements, dynamic combi-natorial chemistry can be used to guide their assembly into ahigh-affinity receptor. Thus, starting from a dynamic combi-natorial library made from two building blocks that containcarboxylate groups as recognition units, a new macrocyclicreceptor for the oligoamine spermine was obtained that bindsthis guest with a dissociation constant of 22 ((1) nM. Water-soluble synthetic receptors for biologically relevant guests withsuch high affinities are extremely rare. Analysis of the host-guest complex by NMR and computer simulations indicates apseudo-rotaxane structure in which spermine is threaded throughthe macrocyclic host and bound through a series of hydrogenbonds and anion-cation interactions. The host was shown tointerfere efficiently with spermine-DNA interactions, being ableto prevent and reverse the binding of spermine to DNA andthereby indirectly control the conformation of DNA; whilespermine can induce a change in the conformation of DNA froma right-handed (B-form) to a left-handed (Z-form) helix, theaddition of the synthetic spermine receptor reverts the DNAback to its original B-form. These results suggest that ourspermine receptor may be a promising lead for the developmentof novel therapeutics or molecular probes that may help unravelthe still poorly understood role of oligoamines in cell biology.

Experimental Section

Materials and Methods. Chemicals were purchased from Aldrichor Fluka. Building block2 was synthesized on the basis of a methoddescribed by Field and Engelhardt (Scheme 1).25 NMR analyses wereperformed using an Advance 500 Bruker instrument equipped with astandard TCI Cryoprobe. Spectra were recorded at 300 K with waterpresaturation during relaxation delay and mixing time.

2,5-Bis(dimethylthiocarbamoyloxy)terephthalic Acid Diethyl Es-ter (8).25 2,5-Dihydroxyterephthalic acid diethyl ester7 (10 g, 39.00mmol) and DABCO (18 g, 157 mmol) were dissolved in dry DMA(100 mL) under a nitrogen atmosphere and cooled to 0°C in an ice

bath. Dimethylthiocarbamoyl chloride (19 g, 157 mmol) dissolved indry DMA (50 mL) was added dropwise under nitrogen. The mixturewas stirred for 16 h at room temperature. The off-white precipitatewas filtered, washed extensively with water (300 mL), and dried undervacuum, yielding compound8 (16 g, 97%).1H NMR (400 MHz,CDCl3): δ 7.73 (s, 2H, CHAr), 4.30 (q,J ) 7.1 Hz, 4H, CH2), 3.46(s, 6H, CH3), 3.40 (s, 6H, CH3), 1.33 (t,J ) 7.1 Hz, 6H, CH3).

2,5-Bis(dimethylthiocarbamoylsulfanyl)terephthalic Acid DiethylEster (9).25 Compound8 (650 mg, 1.51 mmol) was heated undernitrogen at 215°C for 1 h. The brownish mixture was cooled to 70°C,and EtOH (20 mL) was added. Pale brown crystals appeared as thesample was slowly cooled to room temperature. The crystals werefiltered off, yielding compound9 (626 mg, 97%).1H NMR (400 MHz,CDCl3): δ 8.11 (s, 2H, CHAr), 4.34 (q,J ) 7.1 Hz, 4H, CH2), 3.12(s, 6H, CH3), 3.02 (s, 6H, CH3), 1.37 (t,J ) 7.1 Hz, 6H, CH3).

2,5-Dimercaptoterephthalic Acid (2).A solution of compound9(1.3 g, 3.03 mmol) in degassed 1.3 M KOH in EtOH/H2O (1:1, 40mL) was refluxed under an inert atmosphere for 3 h. The reactionmixture was cooled in ice, and concentrated HCl (15 mL) was added.A bright yellow precipitate was formed, filtered, and washed extensivelywith water, yielding compound2 as a yellow solid (663 mg, 95%).1HNMR (400 MHz, CD3OD): δ 8.02 (s, 2H, CHAr).

Dynamic Combinatorial Libraries. In a typical experiment, thethiols (10 mM overall) were suspended in water, and the pH wasadjusted to 8.25 by addition of a 1 M solution of NaOH. Whereappropriate, spermine (2.5 mM) was added. The mixtures were allowedto oxidize and equilibrate by stirring in an open vial at roomtemperature. Evaporated water was replenished every day.

Library Analysis. Analyses were performed using an Agilent 1100series HPLC and Agilent XCT ion-trap mass spectrometer. Solventsand formic acid were acquired from Romil. Analyses were performedusing a reversed phase HPLC column (Agilent C8 Zorbax Eclipse XBD,4.6× 150 mm, 5µm), using an injection volume of 10µL, a flow rateof 1 mL/min, and a gradient (5-95% in 40 min) of acetonitrile in water(both containing 0.1% formic acid) at 303 K. Negative ion mass spectrawere acquired in ultrascan mode using electrospray ionization (dryingtemperature, 350°C; nebulizer pressure, 55 psi; drying gas flow, 12L/min; HV capillary, 4000 V; ICC target, 200 000).

Purification of Receptor 6.Receptor6 was isolated from a dynamic“library” made from thiol2 (10 mM) and spermine (2.5 mM) in waterat pH 8.25. Aliquots of 700µL of this solution were injected onto areversed phase preparative HPLC column (Agilent C8 Zorbax EclipseXBD, 21.2 × 150 mm, 5µm) at 303 K. Using a flow rate of 5 mL/min and a gradient (5-95% over 30 min) of acetonitrile in water (bothcontaining 0.1% formic acid), the receptor was collected from 14 runs.The solvents were removed under reduced pressure keeping thetemperature below 40°C. The last traces of solvent were removed underhigh vacuum for 2-3 h. The crude host was dissolved in 12 mL ofborate buffer (10 mM pH 9.0) and sonicated for 2 min. The resultingsolution was centrifuged and decanted. Hydrochloric acid (3 mL of a2.0 M solution) was added to the supernatant. The resulting suspensionwas centrifuged and the pellet resuspended in dilute hydrochloric acid(30 mL of a 40 mM solution) and centrifuged. The solid (6 mg) wasdried under vacuum overnight. The LC-MS analysis of6 is shown inthe Supporting Information.1H NMR (500 MHz, DMSO-d6): δ 8.34(s, 8H). Anal. Calcd for6‚5H2O: C, 38.32; H, 2.61. Found: C, 38.06;H, 2.76.

Isothermal Titration Calorimetry (ITC). The binding studiesbetween1 and6 were conducted at three different concentrations usingisothermal titration microcalorimetry (MCS-ITC, Microcal LLC,Northampton, MA). Solutions of guest1 (in 3 mM TRIS buffer pH7.4) were titrated into solutions of host6 (in 3 mM TRIS buffer pH7.4). Binding constants and enthalpies of binding were obtained bycurve fitting of the titration data using the one-site binding modelavailable in the Origin 2.9 software as shown in the SupportingInformation. The thermodynamic parameters were averaged over three

(31) The amplitude of the CD signals after addition of spermine and after additionof 6 is somewhat lower than the original value as a result of partialprecipitation of the DNA-spermine complex. This does not affect theconclusions.

Scheme 1. Synthesis of Building Block 2

A R T I C L E S Vial et al.

10256 J. AM. CHEM. SOC. 9 VOL. 128, NO. 31, 2006

experiments (the enthalpy value for run C [see Supporting Information]was discarded due to an inferior signal-to-noise ratio).

DNA Competition Experiments. Analyses were performed usinga Varian Cary 100 Bio UV-visible spectrophotometer and a Jasco J-810spectropolarimeter with 1 cm path-length cells at 298 K. Calf thymusDNA (highly polymerized), poly(dA-dC)‚poly(dG-dT), spermine‚4HCl,and buffers were purchased from Sigma and used without furtherpurification. The solutions were prepared by dilution of concentratedstock solutions of the different species (DNA, spermine‚4HCl, or6) inthe desired buffer (3.0 mM TRIS buffer, pH 7.4, or 3.0 mM TRIS-EDTA buffer, 1.0 mM NaCl, pH 7.4). The final concentration in DNAwas determined by measuring the absorbance at 260 nm (A260) andassuming that 1 absorbance unit corresponds to 0.15 mM in base pairs.Samples were incubated for at least 30 min at 298 K prior to thespectroscopic measurements. The reported UV and CD traces cor-respond exclusively to DNA (contributions from the buffers, spermine,or 6 have been subtracted). Where appropriate, the traces have beencorrected to compensate dilution.

Theoretical Calculations.Calculations were performed using Mac-romodel 8.0 and the AMBER* all atoms force field32 using thecontinuum water solvent treatment. Conformational searches werecarried out using the Monte Carlo Multiple Minimum method (MCMM).For the unbound tetramers, 1000 steps were sufficient to find the 10lowest energy conformations at least twice each. Between 20 000 and40 000 steps were calculated for the complexes; this was sufficient forat least 5 of the lowest 10 conformations to be found at least twice.Subsequent molecular dynamics (MD) simulations, also using the

AMBER* force field with all atom treatment, were run for 20 ns at asimulation temperature of 300 K, with a 1.5 fs time step and 1 ps initialequilibration time. The variation in the average enthalpy (adjusted tothe simulation temperature) during the last 5 ns of each experimentwas determined, and in all cases was found to be less than 2.2 kJ mol-1,implying the simulation length was sufficient. The energies of thespermine, free receptors, and complexes as determined by moleculardynamics are given in the Supporting Information. The interconversionbarriers between P and M forms of the disulfide bonds were estimatedusing the dihedral driving function of Macromodel 8.0. Starting fromthe lowest energy conformer found in the conformational search, thedihedral angle of one disulfide bond was driven, in 1° steps, through360°, in each direction. The lowest energy barrier for the conversionof the empty (P,P,P,P) host into the (P,P,P,M) conformer was found tobe 74 kJ mol-1. In the complex, the energy of interconversion wasdetermined for each disulfide bond, and the lowest barrier was foundto be 83 kJ/mol.

Acknowledgment. We thank J. M. Goodman for advice onthe simulations and J. K. M. Sanders for support. This workwas supported by EPSRC, the European Union (Marie CurieIntra-European Fellowships), and the Royal Society.

Supporting Information Available: COSY spectrum of thecomplex between spermine and6; LC-MS analysis of6; ITCdata for binding of spermine to6; computer simulation results.This material is available free of charge via the Internet athttp://pubs.acs.org.

JA062536B

(32) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.; Lipton, M.;Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C.J. Comput. Chem.1990, 11, 440-467.

Controlling the Biological Effects of Spermine A R T I C L E S

J. AM. CHEM. SOC. 9 VOL. 128, NO. 31, 2006 10257